csv-rest-api

Turn any CSV file into a queryable REST API. Drop CSVs into a directory and they automatically become available as endpoints with filtering, full-text search, sorting, pagination, and field projection — no schema declaration, no database, no boilerplate.

Built for digital transformation: rapidly expose legacy spreadsheet data through a modern API so apps, dashboards, and integrations can consume it without waiting for a database migration project.

Features

• Zero-config: Point it at a folder of CSVs, every file becomes an endpoint at /api/{filename}.
• Auto type-coercion: Strings, ints, floats, and booleans are parsed automatically.
• Filtering: ?department=Engineering
• Full-text search: ?q=alice
• Sorting: ?sort=salary&order=desc
• Pagination: ?page=2&page_size=20
• Field projection: ?fields=name,salary
• Single-record fetch: /api/employees/3 (uses id column or first column)
• Schema introspection: /api/employees/_schema
• Hot reload: Edit a CSV on disk and the change is picked up on the next request.
• Auto-detected delimiters: Comma, semicolon, tab, or pipe.

Install

git clone https://github.com/intruderfr/csv-rest-api.git
cd csv-rest-api
pip install -r requirements.txt

Usage

python csv_rest_api.py --data-dir ./examples --port 8080

Then in another shell:

curl http://localhost:8080/
curl http://localhost:8080/api/employees
curl 'http://localhost:8080/api/employees?department=Engineering&sort=salary&order=desc'
curl 'http://localhost:8080/api/employees?q=alice&fields=name,salary'
curl http://localhost:8080/api/employees/2
curl http://localhost:8080/api/employees/_schema

API Reference

Endpoint	Description
GET /	List available datasets and endpoints
GET /api/{name}	Query a dataset (filters, search, sort, paging, fields)
GET /api/{name}/{id}	Fetch one row by id (or first column)
GET /api/{name}/_schema	Inferred schema for a dataset

Query parameters for GET /api/{name}

Param	Example	Description
q	q=alice	Full-text search across all columns
<column>	department=Sales	Exact-match filter on a column (case-insensitive)
sort	sort=salary	Sort key
order	order=desc	Sort direction (asc or desc, default asc)
page	page=2	Page number, 1-indexed (default 1)
page_size	page_size=20	Rows per page (default 50, max 1000)
fields	fields=id,name	Return only the listed fields

Example response

{
  "dataset": "employees",
  "page": 1,
  "page_size": 50,
  "total": 2,
  "data": [
    {"id": 3, "name": "Carol Davis", "department": "Engineering", "salary": 110000, "active": false},
    {"id": 1, "name": "Alice Johnson", "department": "Engineering", "salary": 95000, "active": true}
  ]
}

Examples

Two sample CSVs ship in examples/:

• employees.csv — staff directory with department, salary, hire date
• products.csv — product catalog with prices and stock

Tests

pip install pytest
pytest

The test suite covers type coercion, filtering, sorting, pagination, projection, and every HTTP route.

CLI flags

Flag	Default	Purpose
--data-dir	./data	Directory of CSV files to serve
--host	127.0.0.1	Bind address
--port	8080	TCP port

Why this exists

Most digital-transformation projects start with a pile of spreadsheets that someone, somewhere, treats as the source of truth. Standing up a "real" database for each one is overkill, and writing bespoke Flask routes per CSV is repetitive. csv-rest-api skips both: drop the file in a folder and you have an API. When the data graduates to a real backend, the API contract is already defined.

License

MIT — see LICENSE.

Author

Aslam Ahamed — Head of IT @ Prestige One Developments, Dubai LinkedIn: https://www.linkedin.com/in/aslam-ahamed/